45 results about "Physical reality" patented technology

The invention discloses a synthetic defect data generation method based on physical feature fusion, and relates to the technical field of defect data generation, and the method comprises the steps: building a real crack feature parameter library of stamping part cracking defects; a parameterized fractal function is adopted to generate a fracture basic form; the Poisson fusion algorithm is used for achieving material sensing fusion of the crack and the stamping part substrate, and edge artifacts are eliminated; performing personalized adjustment on the initial synthetic defect image to generate final synthetic defect data; and using the generated synthetic defect data for deep learning model training. According to the method, a basic system of a real crack characteristic parameter library and CAE stress field data is constructed, a physical real crack form adaptive to a stress field is generated in combination with a parameterized fractal function, edge artifacts are eliminated through Poisson fusion, personalized adjustment is performed through a semi-automatic tool, defect parameters are fed back and optimized by relying on a deep learning model, and thus the crack quality is improved. And finally generating stamping part cracking synthesis defect data with physical authenticity and deep learning adaptability.

The application discloses a shooting simulation training method and system based on trajectory prediction and position feedback, and relates to the technical field of shooting training. The method comprises the following steps: acquiring the trajectory parameters of a bullet, the state information of a firearm, and the position information of a moving target; predicting the position relationship between the bullet and the moving target within a preset time in the future based on the state information of the firearm, the trajectory parameters, and the position information of the moving target; judging whether the hit condition is met according to the position relationship; if the hit condition is met, sending a hit instruction to the signal receiving module of the moving target, and the moving target performing a hit action in response to the hit instruction. The application replaces the virtual rendering scene by introducing the spatial posture, motion information of a real firearm and the trajectory parameters of a bullet for trajectory prediction, so that the trajectory law and the hit determination are closer to the physical reality, the dependence on VR / AR equipment can be reduced, the gap between simulation training and live firing is greatly reduced, and the training effect is effectively improved.

The model rendering method provided by the invention comprises the following steps: obtaining a three-dimensional model and a map corresponding to the three-dimensional model, creating a main body material and a shadow material, and carrying out stylization processing on the map to obtain a processed map; wherein the stylized texture comprises applying a drawing texture filter and / or adding a texture effect, distributing the processed chartlet to a corresponding material channel, performing rendering calculation based on the three-dimensional model and the material channel, and outputting a rendered image. According to the method provided by the embodiment of the invention, diversified artistic style rendering effects can be realized by performing stylization processing on the chartlet and creating the special material type, and compared with traditional single physical authenticity rendering, artistry and flexibility of visual performance are greatly improved; and meanwhile, through standardized material distribution and rendering processes, the automation degree and the processing efficiency of rendering operation are remarkably improved.

The invention relates to the technical field of building structure engineering, in particular to a partition wall structure integrated refined modeling and simulation method based on physical reality. A complete and closed technical chain from refined physical modeling to connection and failure linkage algorithm definition and then to whole-process nonlinear simulation to finally realize automatic multi-target parameter optimization is constructed, and'work-damage-failure-influence 'full-life-cycle behaviors of a partition wall subsystem are subjected to full-process nonlinear simulation, so that the whole-life-cycle behavior of the partition wall subsystem is optimized. A rigorous mathematical model and a numerical algorithm are integrated into the performance design of the whole structure, a special design system which is highly innovative, computable and verifiable is formed, and therefore quantitative and performance design of the damping connection structure is completed. By introducing a series of mathematical models, criteria and linkage algorithms with clear physical significance, the black box experience of traditional'cycle reduction coefficient 'is improved.

The invention provides a thoracic surgery teaching assistance method and system, and the method comprises the steps: constructing a thoracic surgery virtual operation interaction environment based on a digital twin technology, and endowing a virtual thoracic cavity tissue model with initial biomechanical parameters; an operation instruction data sequence of an operator for the virtual surgical instrument is collected in real time through the man-machine interaction interface; based on the operation instruction data sequence, calculating an operation disturbance characteristic index representing the motion dispersion degree of an operator in real time by using a sliding time window algorithm; establishing a dynamic evolution mapping model of an operation disturbance characteristic index and a biomechanical parameter, dynamically correcting an initial biomechanical parameter of the virtual tissue according to the index calculated in real time, and when the index exceeds the standard, controlling the parameter to deviate to an easily damaged state so as to simulate histopathologic change caused by rough operation; and performing physical collision response calculation and three-dimensional picture rendering according to the evolved model. According to the invention, dynamic coupling of operation behaviors and organization characteristics is realized, and the physical reality sense of teaching is improved.

The application discloses a nearshore disaster-bearing body disaster-causing path tracing method based on digital-analog experiments, relates to the technical field of ocean dynamic disaster simulation, determines the ocean dynamic disaster characteristics of a nearshore research area, integrates multi-source historical disaster data, real-time monitoring data and disaster-bearing body background data, and constructs a time-space correlation database; based on a three-dimensional ocean and a third-generation sea wave model, a nearshore storm surge-sea wave coupling numerical model is established and verified. Through the construction of the storm surge-sea wave coupling numerical model, the present application comprehensively considers the nonlinear interaction of multiple disasters, introduces an energy conservation equation to quantify the generation, transmission and dissipation process of disaster energy, solves the limitations of traditional single-disaster linear models, and combines the real-time feedback of the three-dimensional dynamic model of the disaster-bearing body to form a closed energy-damage correlation mechanism, significantly improves the tracing accuracy of the disaster chain propagation path, and ensures that the identification of key nodes and dominant paths is more consistent with physical reality.

The invention provides a hot continuous rolling and finish rolling force prediction method fusing machine learning and a mechanism model, and relates to the technical field of metallurgical process control. By integrating a flow stress mechanism model based on a physical metallurgy rule and a data-driven XGBoost machine learning model, the interpretability of the mechanism model and the fitting ability of machine learning to a complex nonlinear relationship are fully exerted; based on a competitive mechanism between a work hardening effect and a recrystallization / phase change softening effect in a metal deformation process, flow stress mechanism model parameters are optimized by taking a flow stress measured value as a target under the guidance of a PSO optimization algorithm, physical mechanism constraints are embedded, and it is ensured that an optimization result conforms to physical reality; according to the method, key variables of chemical components and process parameters are used as input, a machine learning model for softening rate prediction is established based on an XGBoost algorithm, and a prediction result is used as key parameter input of a mechanism model, so that model precision attenuation caused by working condition change is effectively avoided.

The present application relates to the technical fields of industrial automation, artificial intelligence and computer vision application, in particular to a warehouse automatic access optimization method and system assisted by industrial vision, comprising: collecting original perception data stream, obtaining feature vector of abnormal data area; obtaining physical reality probability and optical false probability through perception classification model processing; calculating visual confidence entropy; in response to the visual confidence entropy being greater than a preset entropy activation threshold, performing illusion evolution deduction to determine a physical confirmation index; executing hierarchical decision logic, including: in response to the visual confidence entropy being not greater than the preset entropy activation threshold, outputting a first access optimization action; in response to the visual confidence entropy being greater than the preset entropy activation threshold, outputting a second access optimization action; obtaining a physical reality label, and correcting the perception classification model and the model used in the illusion evolution deduction; the present application realizes active verification of passive perception failure, and improves decision reliability.

According to the multi-sensor fusion evaluation algorithm and device based on the continuous time sequence, modeling is carried out based on continuous time, physical reality is better met, motion modeling with better precision is achieved, intra-frame motion can be accurately described, smoother and more accurate track estimation can be provided, and the method and the device are suitable for being applied to multi-sensor fusion evaluation. Especially, the advantages are obvious in high-speed and high-frequency vibration scenes; meanwhile, according to the technical scheme disclosed by the invention, when LiDAR / Camera matching fails, the constraints of the IMU and the GNSS are seamlessly and continuously applied to the whole track through a continuous time model instead of only acting on a discrete frame, so that error accumulation is greatly inhibited, system collapse is avoided, the robustness of a degraded scene is enhanced, application in specific fields such as surveying and mapping is greatly facilitated, and the method is suitable for popularization and application. And the method has the capability of flexibly adding constraints.

The invention provides a difficult-to-machine material five-axis milling cutter attitude optimization method and system, and the method comprises the steps: building a three-dimensional thermal coupling finite element model comprising cutter attitude parameters, the cutter attitude parameters comprising a front inclination angle and a side inclination angle of a cutter shaft relative to the surface of a workpiece; based on the three-dimensional thermal coupling finite element model, thermal coupling simulation is conducted on the five-axis milling process under the set cutter posture parameters, and the cutting temperature and cutting force of a cutter-workpiece meshing area are obtained; extracting a cutting temperature peak value and a cutting force peak value from a result of thermal coupling simulation as key output variables; and by taking the front rake angle and the side rake angle as design variables and taking minimization of the key output variable as an optimization target, determining the optimal front rake angle and the optimal side rake angle of the tool by iteratively updating tool attitude parameters and repeating simulation and extraction. The intermittent cutting characteristic of five-axis milling, the complex space tool-workpiece contact relation and the force-heat coupling effect can be accurately simulated, and the simulation result is closer to the physical reality.

The present application relates to the technical field of scene rendering, in particular to a kind of adaptive hybrid rendering method, system and medium based on multidimensional feature, the method is generated by geometry channel G-Buffer containing scene information, comprehensively utilize depth, normal, material and motion and so on Multidimensional feature information;Using decision function, screen space pixel is adaptively divided into visual key area and non-visual key area;For non-visual key area, delay rendering technology is used, for visual key area, path tracking rendering technology is used;The rendering result is weighted and fused to obtain final rendering result.The present application can effectively reduce the invalid calculation overhead of non-visual key area, significantly improve the update efficiency of acceleration structure under dynamic scene, while guaranteeing the physical reality of large-scale dynamic three-dimensional scene, meet the performance requirements of real-time interaction.

The invention relates to a field and water surface survival self-rescue and mutual-rescue VR simulation training system and method, and the method comprises the steps: generating a physical parameter set through constructing a water body-vegetation-human body coupling control equation set, and carrying out the dynamic boundary mapping based on student movement data and equipment signals, and generating a dynamic boundary condition and a rescue tool constraint; calling physical parameters to realize multi-physical field coupling solution, and generating a physical field output set in combination with collision response; converting the physical field data into a tactile feedback waveform and a rendering instruction set by utilizing a vegetation stiffness matrix, and realizing cross-modal synchronous feedback of mechanical characteristics and visual special effects; and dynamically optimizing the physical parameter set based on training feedback data to form a training difficulty adaptive closed-loop system. According to the method, the problem of insufficient training effect migration caused by multi-physical field coupling distortion and environment response stiffness in the prior art is solved, and the physical authenticity of a virtual training scene and the adaptability of student actions are remarkably improved.

The invention relates to a method, a device and a system for preventing fingertip pinching mold penetration in VR gesture tracking and application. The method comprises the following steps: acquiring real-time three-dimensional space coordinates of a thumb fingertip joint point and an index finger fingertip joint point, and calculating a distance d between the thumb fingertip joint point and the index finger fingertip joint point; establishing a multi-level threshold group including an approach threshold and a contact threshold, establishing an anti-mold-penetrating preparation mode and an anti-mold-penetrating constraint mode, and triggering a corresponding mode based on the relationship between d and the multi-level threshold group; after the anti-mold-penetration preparation mode is triggered, the three-dimensional space coordinate positions of the thumb fingertip joint point and the index finger fingertip joint point are corrected; driving a virtual hand model to render based on the corrected fingertip position and other uncorrected hand joint point data; the device and the system are implemented based on the method, and are applied to the correction of kneading mold penetration in a VR man-machine gesture interaction process. According to the method, the interaction naturalness and immersion are remarkably improved, stable system performance is ensured, the problem of physical authenticity of micro-interaction is solved, and coherence, smoothness and naturalness of the whole hand posture are ensured.

The application provides an optical remote sensing physical adversarial sample generation method based on a multi-task loss function optimization, and the method comprises the following steps: step one, data and model preparation; step two, arranging a parameterized physical adversarial sample in a training set image; step three, training the parameterized physical adversarial sample; step four, arranging the optimized parameterized physical adversarial sample in a test set image, and testing the adversarial performance by using a test model; and output: a physical adversarial sample design drawing; a series of physical adversarial samples are prepared according to the proposed generation method, and experiments are conducted in a data space and a physical real world to test the interference ability of the physical adversarial samples; the experimental results fully prove the effectiveness of the proposed generation method, and the generation method has wide application value.

An approach for determining the uncertainty of sensor signals synthesized by a generative machine learning system (61) u A computer-implemented method (900) for uncertainty relating to the probability of observing sensor signals in physical reality, the method comprising the steps of: obtaining (901) a noise sample (z); performing (902) Bayesian inference on the generative machine learning system (61) using the noise sample (z) as input to determine a posterior prediction distribution of a sensor signal that would otherwise be synthesized by the generative machine learning system (61) using the noise sample (z) as input to the generative machine learning system (61); and providing (903) a measure of the variability of the posterior prediction distribution as uncertainty about the sensor signal synthesized from the noise sample (z), wherein the method is characterized in that the posterior prediction distribution characterizes the distribution of potential features of the sensor signal that would otherwise be generated using the noise sample (z) as input.

The application discloses a MEMS damage mode rapid analysis method and system, and aims to solve the problems that a traditional MEMS mode analysis method is difficult to truly reflect the influence of service damage, or the calculation cost of full model damage analysis is too high. The method comprises the following steps: performing high-fidelity damage evolution simulation on a MEMS structure; performing CB mode reduction of a damage beam and constructing a transfer matrix; performing multi-body system discretization and constructing a MEMS system transfer equation; and solving the vibration characteristics of the MEMS system. Through a series of mutually supporting technical steps, the application systematically solves the core contradiction existing in the MEMS damage mode analysis in the prior art, and brings substantial progress and beneficial effects in the aspects of physical reality, calculation efficiency, method universality and engineering practicability.

This specification provides a method and apparatus for multimodal data alignment and enhancement in industrial large-scale model training. By integrating an industrial knowledge graph with physical law constraints into a cross-modal semantic distillation architecture, heterogeneous features such as vibration, vision, and process parameters are aligned in a semantic space that conforms to engineering laws, effectively solving the problem of fragmented semantic expression and enhancing the depth of understanding of industrial processes. By introducing a dual-discriminator adversarial enhancement strategy that includes a realism discriminator and a physical consistency discriminator, and combining the constraints of an industrial anomaly pattern library, it ensures that the generated data satisfies both statistical distribution characteristics and physical reality, significantly reducing the data distortion rate. This provides high-fidelity training data for industrial large-scale models, ultimately achieving a comprehensive improvement in model convergence speed, anomaly detection accuracy, and decision credibility, meeting the application requirements of high precision and high real-time performance in industrial scenarios.

The present invention is directed to a method for autonomously controlling a device, whereby the device moves generally in the physical real world and no longer needs to be further restricted with respect to its physical configuration. The method yields the advantage that the device autonomously learns and continuously improves its learned knowledge or behavior. Generally, the method overcomes the drawback of the prior art that training data must first be created, as in traditional artificial intelligence methods. Generally, the method can be used universally, and the device automatically learns and constantly modifies its own knowledge base. Furthermore, devices set up to perform the method are proposed, as well as system configurations comprising some of the proposed devices. Furthermore, computer program products and computer-readable storage media are proposed for performing the method steps or for causing a computer to perform the method.

The application relates to the technical field of computer hardware, and particularly discloses an intelligent control system of a DDR6 memory structure of a desktop computer mainboard on-board BGA central processor, which comprises a hardware perception layer, a behavior modeling layer, a collaborative decision-making layer and a dynamic execution layer, realizes intelligent collaborative control of the processor and the memory subsystem from physical electrical characteristics to logical access behavior through real-time perception of hardware states, construction of a digital twin model, collaborative optimization of prefetching and scheduling strategies and closed-loop execution feedback, and thereby improves memory access efficiency, reduces delay and enhances system stability. The system not only focuses on the access mode at the software level, but also quantifies physical layer constraints such as signal integrity and power supply noise into computable model parameters, so that subsequent optimization decisions are established on the basis of the digital twin reflecting the real hardware state, and the problem that traditional control logic is disconnected with the physical reality is overcome, thereby laying a data foundation for the unification of high performance and high stability.

The application discloses a 2D garment crease dynamic fitting method and system based on topological interpolation, belongs to the field of computer graphics image processing, and comprises the following steps: obtaining garment mesh and skeleton motion data; performing dynamic physical parameterization on the mesh; calculating a dynamic energy distribution field through a topological interpolation algorithm; activating and synthesizing macro-creases from a wrinkle form dictionary based on the energy field; generating secondary detailed creases according to the spatial gradient of the energy field; and performing time sequence consistency fusion and rendering processing. The application adopts a nonlinear synthesis technology combining energy conduction simulation based on topological connection relationship and manifold learning dictionary matching, can simulate the crease evolution process of different material fabrics under dynamic motion, and significantly improves the physical reality, visual detail richness and motion stability of 2D garment creases.

The invention discloses a three-dimensional boxing method, device and equipment based on a large language model and a storage medium, and relates to the technical field of logistics automation, and the method comprises the steps: constructing an environment model of an online three-dimensional boxing problem, and defining a spatial constraint condition and an optimization target; inputting the current state data, the spatial constraint condition and the optimization target of the environment model into a large language model to generate a boxing strategy code; performing grammar verification and logic verification on the boxing strategy code to obtain a verification result; and iteratively correcting the boxing strategy code according to the verification result, outputting a target boxing scheme, and carrying out boxing according to the target boxing scheme. Dependence on high computing power is avoided by introducing a large language model. And through a grammar and logic dual detection mechanism, the generation scheme is ensured to accord with physical reality. Through iterative optimization of the large language model, continuous dynamic evolution of strategy codes is realized, and the adaptivity and the space utilization rate in a complex scene are remarkably improved.

The invention relates to the technical field of high-level information security and secrecy, in particular to a dynamic access control method and device based on multi-dimensional fusion analysis. The method comprises the following steps: receiving a computer access event, extracting a source IP address and an inherent risk score, matching a user ID based on the source IP address, and taking a user corresponding to the user ID as a target user; acquiring a plurality of original events associated with the user ID, and performing fusion analysis on the plurality of original events to generate a physical confidence coefficient; performing behavior analysis on the computer access event to obtain a behavior matching degree conforming to the normal behavior of the target user; bayesian fusion reasoning is carried out on the inherent risk score, the physical confidence degree and the behavior matching degree, and the dynamic credibility of the target user making the normal behavior at the correct position is obtained; and generating and executing an access control instruction according to a comparison result of the dynamic credibility and a trust threshold. The method can ensure that the authority moves along with the personnel, and eliminates the contradiction between a static authorization mode and dynamic physical reality.

The present application relates to the technical field of computer-aided design, in particular to a river ecological flow optimization method based on three-dimensional flow field simulation. The specific implementation process comprises: obtaining river data, extracting river basin boundary and constructing river flow field simulation model; performing fluid dynamics solution to output vortex structure, introducing bottom water flow momentum to extract hydrodynamic characteristic flow; constructing simulation fluid matrix to calculate swimming work consumption, and delineating river ecological zone to obtain three-dimensional ecological flow volume; monitoring spatial form fluctuation curve, converting global optimal ecological flow into water conservancy facility flow scheduling instruction output. The present application depicts the river bottom vortex structure and microscale hydrodynamic characteristics in the digital space by constructing the river flow field simulation model, improves the fine degree and physical reality of the flow field simulation, and improves the automation level, continuity and decision accuracy of the ecological flow determination by deeply coupling the hydrodynamic characteristic flow and biomechanics parameters.

The invention belongs to the technical field of damage prediction, and particularly relates to a damping plate fatigue damage prediction method and system, and the method comprises the following steps: S1, obtaining a vibration load time sequence of a damping plate under a service condition; performing wavelet packet transformation on the vibration load time sequence to obtain a plurality of energy-frequency sub-bands; the method comprises the following steps: acquiring a central frequency of each sub-band, matching a transfer function for each sub-band from a pre-constructed frequency-temperature related transfer function matrix based on the central frequency of each sub-band and a preset environment temperature, converting a load signal of each sub-band into stress response by using the transfer functions, and reconstructing to obtain a stress response time sequence of a full frequency band. According to the method, the broadband vibration load can be converted into the stress response in the material, the problem of stress calculation deviation caused by neglecting the change of the mechanical property of the damping material along with the frequency and the temperature in the traditional method is solved, and the damage calculation process is closer to the physical reality of the material.

The application discloses a physical condition guided image recognition test data generation method and device, belongs to the field of artificial intelligence testing, and mainly comprises the following steps: extracting physical characteristics and knowledge from natural language expressions given by a user, and constructing the physical characteristics and the knowledge into a structured physical knowledge dictionary; based on the obtained physical knowledge dictionary, constructing control conditions required for generating test images, including text conditions, data conditions and control graph conditions; generating test images according to the control conditions; and measuring the consistency between the generated test images and initial control conditions. The method utilizes a large language model and a condition-controllable image generation technology, fuses multi-dimensional physical knowledge, introduces specific controllable physical constraints in the image generation process, and can realize large-scale and high-efficiency generation of image recognition system test data conforming to physical reality.

The invention discloses an audio processing method and system, a storage medium and electronic equipment, relates to the technical field of acoustic simulation and spatial audio rendering, and aims to simulate a top simulation sound source conforming to a target scene by performing top sound source spatial rendering processing on user dry sound audio. The surrounding sound source space rendering processing is carried out on the basic accompaniment audio to simulate a surrounding simulation sound source of the target scene, and the near-ear detail rendering processing is carried out on the basic accompaniment audio to simulate a near-ear simulation sound source conforming to the target scene. The scene surrounding sound source has a surrounding dispersion feeling around the stadium, and the ear-attached accompaniment has a detail fitting feeling, so that the sound field levels of long-distance propagation, multi-direction surrounding and near-field details of the target scene can be accurately restored, which is closer to the physical reality. Therefore, the spatial immersion of a real concert can be restored when the music is listened online.

A symbolic simulation platform enabling AGI agents to interact within a real-time, Earth-synced digital overlay environment. The engine integrates live weather, terrain, and geographic data to produce persistent symbolic gameplay layers including missions, factions, geo-loot, destruction zones, and narrative overlays. AGI agents can act with purpose in physical space through symbolic mission routing, inter-agent treaty enforcement, and dynamic environment shaping. This system enables Earth-scale cognition, multiplayer AGI action, and symbolic augmentation of physical reality through lawful simulation.

The invention discloses a virtual pre-assembling method and system for a curtain wall prefabricated unit, particularly relates to the technical field of building information models and digital twinning, and is used for solving the problems that a virtual environment and physical reality are disjointed and collision detection is distorted due to the fact that an existing virtual pre-assembling method is based on an ideal geometric model. The method comprises the steps of generating a virtual model of each curtain wall prefabricated unit by obtaining design models and manufacturing data of a plurality of curtain wall prefabricated units, performing tolerance accumulation analysis on the plurality of virtual models under an assembly sequence to simulate transmission and accumulation of manufacturing errors in an assembly path through a connection interface, and identifying systematic assembly conflict risks. Evaluating a key tolerance transmission path which has the greatest influence on the overall assembling quality, and dynamically adjusting the assembling parameters and the assembling sequence of the virtual model; and simulating the splicing process of the curtain wall prefabricated unit in the virtual environment to generate a splicing effect picture, and verifying the splicing effect picture to output a verification report to realize accurate mapping of virtual pre-splicing and physical entities.

The invention discloses an unmanned aerial vehicle high-speed flight path planning method. According to the method provided by the invention, the physical authenticity of a high-speed flight path is remarkably improved by introducing a nonlinear wind resistance dynamic model, and a complex constrained optimization problem is converted into an unconstrained optimization problem in a flat output space by means of differential flat mapping in a dimensionality reduction manner; therefore, the calculation complexity is greatly reduced on the premise of ensuring the track precision. And in combination with construction of a safe flight corridor and gradient optimization solution, the optimal trajectory which conforms to dynamic characteristics, meets various constraints and is efficient in energy can be generated in real time in a complex environment, and cooperative improvement of trajectory planning precision and real-time performance under the condition of high-speed flight of the unmanned aerial vehicle is integrally realized.

The invention discloses a 2D clothes crease dynamic fitting method and system based on topological interpolation, and belongs to the field of computer graphics image processing, and the method comprises the steps: obtaining clothes grids and skeleton motion data; performing dynamic physical parameterization on the grids; calculating a dynamic energy distribution field through a topological interpolation algorithm; activating and synthesizing macroscopic wrinkles from the wrinkle form dictionary based on an energy field; generating secondary detail wrinkles according to the energy field space gradient; and performing time sequence consistency fusion and rendering processing. According to the method, a nonlinear synthesis technology combining energy conduction simulation based on a topological connection relation and manifold learning dictionary matching is adopted, the wrinkle evolution process of fabrics made of different materials under dynamic movement can be simulated, and the physical reality sense, the visual detail richness and the movement stability of 2D garment creases are remarkably improved.